You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Describe the issue
When Sablier scales up a previously scaled down deployment, it always scales it to 1 pod, and does not return the deployment back to the value of replicas specified in the original YAML.
Context
Sablier version: All
Provider: Kubernetes
Reverse proxy: N/A
Sablier running inside a container? Yes
Expected behavior
It should be possible to scale back up to multiple replicas when scaling back up from zero. This could either be by capturing the original spec.replicas for the deployment and using that, or introducing a new sablier.replicas label.
Additional context
This would allow Sablier to be used in scenarios where workloads should both scale to zero when not in use and scale to a highly available multi-instance deployment when in use.
I could try to contribute this as a PR if you are open for this......
The text was updated successfully, but these errors were encountered:
Do you know how I can retrieve the original replicas ?
What happens if the original replicas is 0 ?
I can add this fairly easily with groups, however the API using names query param would not support it properly, or it would scale all mentioned workloads to this specific replicas.
Describe the issue
When Sablier scales up a previously scaled down deployment, it always scales it to 1 pod, and does not return the deployment back to the value of
replicas
specified in the original YAML.Context
Expected behavior
It should be possible to scale back up to multiple replicas when scaling back up from zero. This could either be by capturing the original
spec.replicas
for the deployment and using that, or introducing a newsablier.replicas
label.Additional context
This would allow Sablier to be used in scenarios where workloads should both scale to zero when not in use and scale to a highly available multi-instance deployment when in use.
I could try to contribute this as a PR if you are open for this......
The text was updated successfully, but these errors were encountered: